SIDDHARTH JAIN

• Collaborated with stakeholders to define business requirements for analytics projects.
• Managed multiple projects simultaneously while ensuring timely delivery according to requirements through the use of MLFlow, ClickUp and JIRA.
• Ensured accuracy of results by validating model assumptions against real-world scenarios or domain knowledge.
• R&D and Automation: Conducted research and development to automate feature engineering, particularly focusing on custom deep learning models

Project Leadership:

• Cat Identification System: Developed a cat identification system for smart bowls using YOLO for frame selection, EfficientNet for embedding generation, and clustering for classification
• Pet Activity Prediction and Health Monitoring System: Utilized weight sensor data to predict pet elimination activities with Random Forest and XGBoost, conducting in-depth analysis of multi-day behavior features to assess pet health status
• Canine Activity Prediction: Analyzed accelerometer and gyroscope data from smart collars to forecast canine activities using a combination of CNN and LSTM models compatible with edge computing
• MVP Phase and Data Pipeline: Tackled real-time issues during the MVP phase and gained insights into the data pipeline, Power BI, managing AI model training, data analysis, feature engineering, and real-time prediction projects

• Power Forecasting: Led a team of 12 in developing wind, solar, price, and load power forecasting models, addressing operational issues
• Wind Power: Developed custom models based on domain expertise.
• Solar Power: Applied computer vision and deep learning (TensorFlow) on satellite images for solar radiation forecasting
• Load Forecasting: Used ML algorithms on historical data for day-ahead demand forecasting
• Price Forecasting: Applied Holt-Winter and SARIMAX models for day-ahead power price prediction
• Unit Commitment: Developed day-ahead operation schedules using Mixed Integer Linear Programming with GAMS, considering various constraints and decision variables

• Sales Forecasting: Forecasted monthly sales for pharmacy products, optimizing supply chain using ARIMA, Holt-Winter, and sensitivity analysis
• Loan Application Automation: Automated client verification for loan approvals using IBM Watson's NLP capabilities
• Price Elasticity Analysis: Determined price elasticity of demand using historical sales data